Pruned-ADAPT-VQE: Compacting Molecular Ansätze by Removing Irrelevant Operators

Area of Science:

• Quantum computing
• Quantum chemistry
• Computational physics

Background:

• The adaptive derivative-assembled problem-tailored variational quantum eigensolver (ADAPT-VQE) is a key quantum algorithm for electronic structure calculations.
• ADAPT-VQE's adaptive operator selection can lead to inefficient ansätze with redundant or near-zero parameter operators.
• Identifying and addressing operator redundancy is crucial for optimizing ADAPT-VQE performance.

Purpose of the Study:

• To propose an automated, cost-free method for refining ADAPT-VQE ansätze.
• To reduce ansatz size and improve convergence speed in ADAPT-VQE calculations.
• To enhance the efficiency and reliability of quantum electronic structure computations.

Main Methods:

• Developed a post-optimization refinement technique to identify and remove unnecessary operators.
• Introduced an evaluation function considering operator parameter value and position.
• Implemented a dynamic thresholding mechanism for efficient convergence.
• Applied the method to various molecular systems.

Main Results:

• Demonstrated significant reduction in ADAPT-VQE ansatz size.
• Observed accelerated convergence, especially in systems with flat energy landscapes.
• Showcased consistent improvement or maintenance of ADAPT-VQE performance.
• Confirmed minimal additional computational cost for the refinement process.

Conclusions:

• The proposed refinement method effectively removes redundant operators from ADAPT-VQE ansätze.
• This technique enhances computational efficiency and convergence for quantum electronic structure calculations.
• The approach offers a practical solution for optimizing ADAPT-VQE performance in quantum computing applications.